How Minut works to protect privacyA post by Minut CEO, Nils Mattisson
Monitoring done differently
From the beginning, Minut has approached home security differently. Privacy for our users and their homes has always been central to our approach, and our smart home alarm reflects this concern across every aspect of its design.
Six years ago, when we first began to think about home security, options on the market were limited to the invasive option of wifi-connected security cameras, and the considerably more expensive traditional home alarm systems. My co-founders and I wondered why there was no middle ground, where was the option for those who did not want, need, or could not afford an expensive full security system, but did not feel comfortable cameras in their homes? We considered ourselves amongst this underserved group, and as engineers often do, decided we could design something better: an affordable security system that could provide the peace of mind of knowing that everything is safe at home, without sacrificing any privacy.
This is our vision for a gentler approach to home security that places privacy and protection at its core, and is accessible to those for whom home security has always been out of reach. With this in mind, we have built a system that does not include cameras and uses edge computing meaning data is processed immediately and never stored.
Privacy in the Smart Home
Home is our most private space and the area where we most crave a sense of security. Home simply isn’t home without it, but as more and more everyday technologies become ‘smart’, companies are also gaining access to some of our most intimate, personal data through the events inside our homes. The internet of things (IoT) has brought ‘smart’ technology to even the most banal of household items, and changed the way we think about tech in the home. IoT devices have quickly become a fixture for households, and the industry shows no signs of slowing down. Worldwide management consulting firm McKinsey estimates that the number of connected smart home devices will surpass 75 billion by 2025.
This exploding industry has brought a new level of connectivity, efficiency, and convenience into our lives and homes, but as these connected technologies continue to grow in prevalence and popularity, concerns around the ways in which companies can utilize this incredibly personal data have also arisen. At its best, the smart home can help streamline and simplify our daily lives, but these devices can also serve as a source of highly lucrative data from inside our homes.
Pam Dixon, the executive director of the World Privacy Forum explained the potential risks and implications for the smart home, cautioning that, “Many of these [smart devices], when taken by themselves, do not represent significant privacy risks, the larger concern is where the data goes once it’s stored on a company’s servers. How long is it kept? Can users delete it? Will a third-party company have access to it, and if so, what will it be used for? If these answers to these questions aren’t clear (Minut terms of service), people’s personal data might be accessible to all sorts of businesses — advertisers, insurers–without their knowledge”. Dixon noted that these questions are of particular interest to the largest tech companies, “If there’s a single platform that knows where you’re driving and your thermostat temperature and your security camera installation and other aspects of your life, I do think that warrants a higher degree of scrutiny.” The scrutiny that Dixon calls for, has been answered to some degree by GDPR regulations within the EU, but for users outside of Europe, particularly in the U.S., the companies behind these devices remain largely autonomous in deciding how their customer’s data is utilised, and how much of this data collection must be disclosed to users.
Moving to the Edge of the Network for a Truly Intelligent Smart Home
Much of the ambiguity surrounding data collection in the smart home centers around what happens to personal data once it has been transmitted for storage on remote cloud servers.
In the process of developing our alarm, we reached the conclusion that the only way to truly guarantee the privacy of personal data was to avoid collecting data as much as possible. In practice, our ‘alarm’ is actually a very small computer equipped with an array of environmental sensors and a powerful processor. This processor applies advanced algorithms to analyse noise, motion, and temperature data in real-time without the need to record sounds or transmit raw data recordings to the cloud for processing. By combining the inputs of its environmental sensors with powerful machine learning, the alarm is able to process and analyse data on the device itself, ensuring no raw audio from the home is ever recorded or sent to the cloud for storage. Only when the alarm senses a potentially critical event (such as glass breaking or a sudden sharp increase in temperature) is a ‘fingerprint’ of the potential anomaly generated and sent to the cloud for a ‘second opinion’ of sorts, where it is analysed by even more powerful algorithms before being sent to the user.
The prevalence of IoT devices in the space of the smart home, has given rise to a slew of connected technologies all claiming to be ‘smart’. However a great deal of these so-called ‘smart’ devices lack any true intelligence on their own, as they simply a vessel for sensors that upload all data to remote cloud servers for analysis.
The Cloud first made it possible to access computing resources anywhere, regardless of individual device location. However, as more and more devices become connected through the Internet of Things, the privacy of data stored within the cloud has, for many, become too much of a question mark. Once data from the smart home is stored remotely, it is nearly impossible to definitively determine just how closed off or private this data really is, and exactly who has access to it.
As the computational power of individual devices continues to increase, the mechanisms for processing and analysing data are also moving towards a more localised approach that harnesses the computational potential of individual devices. This localised model of computation, commonly known as ‘edge processing’, simply relocates the computational power of the cloud to the edge of the network, in this case your own home network. By bringing the cloud’s processing power and tools for data analytics into the home network, users can ensure that no data from inside their homes ever needs to leave to be processed. When combined with AI and machine learning, this local processing gives devices themselves the ability to process data in real time without ever having to transfer any data to the cloud.
Moving computational power to the edge of the network makes for a truly intelligent smart home, and opens up new possibilities in terms of data protection and privacy. However, for all the innovation and potential of this alternate model of computing, most of us have not widely encountered this technology, and most of the largest players in the tech world have been hesitant at best in adopting and exploring local approaches to data analytics. Collecting user data is far too lucrative of a venture to sacrifice, and the sustainability of these company’s business model relies heavily upon auctioning off user data. This is not to say that the applications of this data are always nefarious, but in regards to data generated inside the home, the lack of transparency surrounding the ways in which this data is utilised, and by whom is particularly unsettling.
As a security system with privacy at its core, edge processing has been a core component of our approach to user data privacy. My team and I utilised this localised approach to data analytics at a time when the industry was still focused almost entirely on cloud-based operations within the smart home. This approach has allowed us to ensure the privacy of our user’s homes, maintain the affordability of our system, and create a truly intelligent security system with privacy as a foundation rather than surveillance.Nils Mattisson, Minut CEO